This is incorrect. It would be the tech equivalent of making it illegal to sell a device that FORCES other peoples rental cars to all play La Cucaracha without their permission. Computer fraud act is going to kick in.
You've got to admit that such a device would be funny as heck.
But it really isn't equivalent to forcing other people's cars to play the tune. Rather, it is equivalent to allowing you, as the driver of the car, to force other nearby drivers to *listen* to the tune. The fact that their cars are vibrating sympathetically in response to external sound waves sent out by your car is merely a side effect of the way the cars and the medium in between them are designed, in much the same way that the misbehavior of the other people's games in response to your game doing something it should not be allowed to do is a side effect of the way the game and the server in between is designed.
They're not hacking other people's devices. They are merely cheating in ways that the server isn't catching, and in ways that cause things to happen that aren't supposed to be able to happen, resulting in situations that other people's devices are unable to help them get out of (because they shouldn't be possible). That's subtly different; the other users' devices are not being forced to do something; they are simply failing to reject invalid input. Similarly, the servers are not being forced to do something; they are simply failing to reject invalid input.
CFAA *might* kick in — it isn't open and shut — but it is clearly not a copyright issue.
The thing is, the right of someone to hack and enjoy the game the way they want ends where it affects other people.
Agreed. But the fact that it annoys other people still doesn't make it a copyright law issue. It's a combination of a contract law issue and a technical issue.
More to the point, it isn't an issue for the person creating the hacking tools, period, because those people likely aren't the ones bound by that contract. Rather it's an issue for the griefers who are using the tools in ways that are sufficiently inappropriate that they should be banned for doing it, it's an issue for the game company that should detect that abuse and ban them, and it's an issue for the innocent gamers caught in the middle by that company's apparent inability to do so.
None of those points support this being treated as a copyright issue in any way. Illegal, perhaps. Subject to lawsuit on other grounds, probably. But a copyright violation? Definitely not.
The "always online, server hosted" game model changes the situation for the same reason companies love it. You are not, by any interpretation of the agreements, buying a game.
Makes no difference. This is the tech equivalent of making it illegal to sell horns that play "La Cucaracha" that are designed to magnetically mount to rental cars because it would cause irreparable harm to the reputation of the rental car company if cars with Hertz license plate brackets sounded like Mexican ice cream trucks (*) or whatever.
* Any implication that Mexican ice cream trucks actually play La Cucaracha is entirely unintentional, just so we're clear.
At best you are purchasing "unlimited access to game servers contingent on respecting the terms of play.
Correct. And to the extent that they can use technical measures to detect abuse and ban people who use these hacks, that is well within their right.
However, a legal bright line is crossed when they attempt to use copyright as a means of regulating what you can do with hardware that you own, or as a means of regulating the sale of products that merely enable the user to modify a copyrighted work. Even in the prior cases, the copyrighted work was licensed, not sold, and the fact that the newer games use the company's servers does not change the fundamental nature of the argument in any meaningful way, because it does not change the fundamental nature of the act in question (the user modifying a game) or the technology in question (a device that helps the user do so) in any meaningful way.
What the company might be able to do is allow the technology to be sold, then sue the seller for damage to their reputation under the computer fraud and abuse act. This is, of course, entirely orthogonal to the copyright issue, which is settled law.
Construction is only part of it. Yes, they often build commercial buildings cheaply, under the assumption that after twenty or thirty years, they will be so dated that nobody will want to rent them. But the bigger problem is that many building owners do minimal maintenance, under the assumption that any maintenance beyond what is critical — maintenance that could extend the buildings' life for decades — would probably not be cost-effective, because they'll be tearing them down to build something new after two to four decades anyway.
Mind you, I do think that $200 is a gross underestimate, but if the city valued it at a billion dollars, that's a laughable overestimate. There's no way you'd get anywhere close to that for a bunch of forty-year-old buildings, no matter how much history they might have.
Oops. I have to correct myself. The IL campus is only 25 years old, not 40. So it still has some life left in it. It's the Mariani buildings that are almost 40 years old.
Considering the value of the land (and yes, ground that is leased has value) you are saying an 850,000 SF building in one of the highest rent places in the world should be worth $235 per square foot. And you are basing this on - the relocation costs for Apple?
I'm basing this on the fact that buildings depreciate over 40 years for a reason. That's the expected life of the building. The buildings in question are now forty years old. So from a value perspective, these buildings are bulldozer fodder. They're not worth anything per square foot at that age. Anybody who buys them will be buying them solely for the right to lease the land for ~55 years. Mind you, if that lease is way below market value (and it may be), then the lease might have significant value, but either way, that value isn't in the building, but rather in the land use rights.
Commercial buildings are depreciated over 39 years.
If this is their actual headquarters (Infinite Loop), then based on forty-year depreciation, a forty-year-old building is worth basically zero, so $200 seems about right. If this is Apple Park, then yeah, this is fraud.
Actually, it isn't entirely insane, assuming this is talking about the Infinite Loop campus. Apple does not own the land under the building. It is owned by Sobrato, the development company on the corner. Apple merely has a 100-year lease on it. So if Apple decided to sell the buildings, absent some agreement by the landowner to allow the lease to be transferred, they would not be able to do so. Arguably, then, the buildings have zero value beyond what the landowner is willing to give them for them.
And even if the landowner agreed to a lease transfer, the buildings would still only have value if somebody else wants them as-is. The problem is, IL1's lobby area had serious mold problems fifteen years ago, particularly on the upper floors. I can't imagine it has gotten any better since then. If Apple ever left, there's a nonzero chance that the next company would decide to tear those buildings down rather than fix them.
It could well be that the expected amount of money that they could get for transferring the lease would not significantly exceed the amount of money they would have to spend bulldozing the old campus to make it ready for whatever company would take it over.
Mind you, I do think that $200 is a gross underestimate, but if the city valued it at a billion dollars, that's a laughable overestimate. There's no way you'd get anywhere close to that for a bunch of forty-year-old buildings, no matter how much history they might have.
And given that the original 100-year lease is almost halfway up, and at the end of that 100 years, the buildings potentially become a giant teardown liability unless the lessee is willing to move them somewhere else, the value of those buildings is at least arguably going to go *negative* at some point.
So really, the only reasonable way to value the property is to determine how much Apple would have to pay to move the employees that are currently in the Loop to other, rented office space, multiplied over the expected remaining life of the building — maybe ten years on the high side. If we assume that they stopped doubling and tripling up in IL offices after Apple Park opened, that's probably only a couple of thousand people. And assume that any new space would be high-density, open plan office space at 175 square feet per employee. Assuming about $4 per square foot per month times ten years, that's about $168 million. At $8 per square foot for demolition times 850,000 square feet, the buildings themselves are a $6.8 million liability, so its value is really closer to $160 million. Seems like a much more plausible number than a billion, which would basically require assuming that Apple will continue using those buildings as-is for the remainder of the hundred-year lease.
The basic problem is that they are hoping they can teach the neural net to drive just based on what it can see, where as every successful self driving project has worked on the basis of building up a 3D model of the world around the car and navigating through it. Lidar is immensely helpful for doing that. In theory a camera might work, but getting a neural network to recognize geometry and spit it out in a format that can be combined into a 3D model is science fiction level strong AI.
Not really. Lidar basically just gives you a depth map. You can get the same thing from a pair of cameras, which Teslas have, at least on the front (the only direction that really matters that much). That doesn't require sci-fi-level strong AI; that's literally what the iPhone does every time you take a photo in portrait mode. It's basically a glorified version of an MPEG encoder's P-frame motion vector computation. It doesn't even require ML, much less strong AI. It is just a bunch of motion vector computation between two reference frames.
The Mendocino Complex fires were reported on July 27, when they burned about 4000 acres. Suppose you decided the very next day to send your bombers, and they arrived on the 29th. By then they'd be facing a fire of well over 50,000 acres, because the wildfire had spread, well, like wildfire.
Just tell them that there's black gold (oil) in them thar hills, and they'll be ready to bomb it in under an hour.:-D
Slashdot posted the bug this morning. If I'm reading it right, the bug was reopened at 10:59 last night. So unless there's time travel involved or something....:-D
And it's not working; they just announced a new hardware revision that has to be retrofitted to existing cars.
I think most of us were pretty certain a long time ago that it would require a faster computer than what they were using. That said, if they're really able to get their machine learning model to process 2,000 frames per second with the new hardware, up from 200, they went from limping along at 25 fps per camera to ostensibly processing 250 fps per camera, which finally exceeds the actual frame rate of the camera hardware (60 fps).
I wouldn't call that "not working". I would call that grossly underestimating the compute power required for self-driving. Chances are, they'd have encountered the same problem if they had used LIDAR like Waymo. It isn't a panacea; it's just another data source.
Theres a lot of reasons to dislike Monsanto, but this probably isn't one of them. Glycophosphate isn't new, it has been used in tremendous quantity all over the US for more than half a century.
It was introduced in 1974, but its popularity really took off in the 1980s. Interestingly, cancer incidence in the United States has been on a steady climb in the U.S. since about 1980. That increased incidence could entirely be a side effect of better detection, or Monsanto could be responsible for one of the biggest public health crises in our nation's history, or anything in between.
...for every one person like that there are a thousand who would like to work for Google.
And probably at least one who has a different opinion on whatever specific issue that person is complaining about. After all, although there are some moral absolutes, there are a lot more situations where different sets of morals conflict, such as the conflict between getting self-driving tech onto the roads sooner to save lives when drivers are half asleep versus delaying it until it is better than those drivers when they are awake.
I grew up with cats, and if you can't get the cat back in the bag it is usually because you're holding it wrong.
If you're holding either one, you're doing it wrong. Step 1: Place the bag in such a way that the bottom lies flat, with the sides piled up around it. Step 2: Wait for the cat to think that it is a box and squeeze between the vertical walls. Step 3: Raise the sides of the bag in a quick vertical motion.
If this does not work, try putting a small amount of cat food on the bottom of the bag.
Abandonware was that, abandoned software. Now that it's not abandoned anymore, time to move on.
No, they're still abandoned. Just as a mother screaming "Don't you touch my child" from a thousand miles away when she hears a bear over a cell phone does not qualify as protection, neither does sending out C&Ds for copying a game that is unavailable commercially qualify as proof of non-abandonment.
I fully agree with the folks who say that there should be a maximum rest period for copyrighted works. If companies want to put something temporarily on hold to allow for demand to build up enough to re-release it, that's one thing, but putting distribution on hold indefinitely is quite another, and runs almost directly contrary to the copyright act's stated purpose of encouraging the creation of new works. If a work has been off the market for ten years, it should automatically be ceded to the public domain.
So no, don't move on. Write your congresspeople. Tell them that you think the copyright act needs to be amended to make redistribution of long-term-unavailable works per se legal. Tell them that statutory limits on how long a rights owner can prevent a work from being available on the market can result in opportunities for new businesses to form and make money off of rights that otherwise presumptively will never again be exercised in any meaningful way. Tell them that these limits contribute to the ability of others to create derivative works that can actually create long-term benefits for the original rightsholder. And tell them that allowing a company to simply bury a work and prevent any new people from ever being able to see it through abusing copyright is just plain dumb.
If enough people do that, we might just make something happen.
This. I predict a week later, these will get auto-deleted by a bot for being insufficiently noteworthy. And thus will begin the first AI-powered edit war.
That's what the nighttime detergent spray is for. It washes the oil off their feathers, and then they freeze to death. Then, you just have to deal with the yard full of dead birds.
ESPN almost certainly doesn't subsidize Disney. They pay almost $2 billion for pro football alone, which is likely more than the total budget for all Disney Channel shows put together.
That said, other NBC-Universal properties might be subsidizing Sci-Fi somewhat. And certainly, all those crappy shopping networks subsidize Sci-Fi by paying cable providers to carry them. (I refuse to call it SyFy.)
What I want to know is why do we need channel bundles anymore?
Mainly so that all the parents with little kids can pay fees for Disney's mostly low-cost programming, to subsidize ESPN so that people who watch sports don't have to pay $30 a game.
You seem to have momentarily forgotten how prices are determined - by supply vs demand. If people want more homes like the did here in Dallas, builders build more homes, like they did here in Dallas. If the wood and other materials to build the condo costs $30K and the labor costs $12K, the builder can make a nice profit selling it for $55K. That's what they've done in Dallas.
Dallas isn't land-locked. It can expand almost infinitely. The Bay Area has mountains on basically all sides, and the land is all in active use. Builders *are* building homes as fast as they can, but it is hard, because before they can build, they have to buy something smaller and tear it down. If the land under that condo costs 12 million dollars and can only realistically handle 30 units, the builder would lose almost half a million dollars selling it for $55K. That's what's happening in the Bay Area. If you want to live in Sacramento, you can probably get a condo for $55K. You'll just have to commute for two hours each way every day down a highway that goes through one of only a handful of mountain passes into the Bay Area.
Under prop 13, cities and counties aren't allowed to get much property tax revenue if they allow land to be used for housing; they have to zone it commercial or office so they get much more property tax revenue from a given parcel of land. So each city is very strongly incentived to approve a shopping center being built on a particular piece of land rather than an apartment complex.
First, you're wrong. Commercial property is covered under Prop 13, too. This is one of the serious problems with that law that many Californians would like to fix, because it would bring in a lot of much-needed revenue from big businesses.
Second, most of the new construction is mixed-use these days, with shopping on the first floor and housing on the upper floors. You're right that in some areas (San Francisco in particular), efforts to build more housing have sometimes been thwarted by government idiocy, but the bulk of the problem is actually caused by allowing a few businesses to become too big to fail.
Basically, there are three big employers: Apple, Google, and Facebook. These three employers have a very disproportionate impact on hiring.
Apple built in Cupertino back when land was cheap.
Google bought the SGI headquarters in Mountain view back when land was cheap.
Facebook built/bought in the swampland part of Menlo Park because nobody wanted to locate there, and land was relatively cheap.
The problem is that they are all close together. With no traffic, it takes about 14 minutes to get from Apple to Google, and ten minutes from there to Facebook. More importantly, traffic going to Google and Facebook are basically using highway 101, taking nearly consecutive exits. The peninsula is a nightmare as a result, traffic-wise. And more importantly, they are all three towards the northwest corner of the South Bay, relatively speaking. So any housing built towards the southeast (the only remaining direction for expansion, down a narrow valley) requires everyone to travel the same highway for most of their trip.
Because they are so close together, the housing within driving distance of those companies is untenable. And there is almost no business construction in areas like Gilroy where housing is more affordable, which otherwise could create reverse commutes and make the whole traffic situation more sane, because business leaders think that somehow they won't be able to get people who live in the South Bay to commute out of the South Bay. Fixing this would help somewhat.
Finally, as previously noted, the Bay Area is largely land-locked, with mountains in nearly every direction. So unlike other cities that just expand over time, the Bay Area really can't.
CEQA is another California law that adds an average of 2 1/2 years to each construction projec
Slashdot Mirror
On this side of the pond, it's the CFAA.
You've got to admit that such a device would be funny as heck.
But it really isn't equivalent to forcing other people's cars to play the tune. Rather, it is equivalent to allowing you, as the driver of the car, to force other nearby drivers to *listen* to the tune. The fact that their cars are vibrating sympathetically in response to external sound waves sent out by your car is merely a side effect of the way the cars and the medium in between them are designed, in much the same way that the misbehavior of the other people's games in response to your game doing something it should not be allowed to do is a side effect of the way the game and the server in between is designed.
They're not hacking other people's devices. They are merely cheating in ways that the server isn't catching, and in ways that cause things to happen that aren't supposed to be able to happen, resulting in situations that other people's devices are unable to help them get out of (because they shouldn't be possible). That's subtly different; the other users' devices are not being forced to do something; they are simply failing to reject invalid input. Similarly, the servers are not being forced to do something; they are simply failing to reject invalid input.
CFAA *might* kick in — it isn't open and shut — but it is clearly not a copyright issue.
Agreed. But the fact that it annoys other people still doesn't make it a copyright law issue. It's a combination of a contract law issue and a technical issue.
More to the point, it isn't an issue for the person creating the hacking tools, period, because those people likely aren't the ones bound by that contract. Rather it's an issue for the griefers who are using the tools in ways that are sufficiently inappropriate that they should be banned for doing it, it's an issue for the game company that should detect that abuse and ban them, and it's an issue for the innocent gamers caught in the middle by that company's apparent inability to do so.
None of those points support this being treated as a copyright issue in any way. Illegal, perhaps. Subject to lawsuit on other grounds, probably. But a copyright violation? Definitely not.
Makes no difference. This is the tech equivalent of making it illegal to sell horns that play "La Cucaracha" that are designed to magnetically mount to rental cars because it would cause irreparable harm to the reputation of the rental car company if cars with Hertz license plate brackets sounded like Mexican ice cream trucks (*) or whatever.
* Any implication that Mexican ice cream trucks actually play La Cucaracha is entirely unintentional, just so we're clear.
Correct. And to the extent that they can use technical measures to detect abuse and ban people who use these hacks, that is well within their right.
However, a legal bright line is crossed when they attempt to use copyright as a means of regulating what you can do with hardware that you own, or as a means of regulating the sale of products that merely enable the user to modify a copyrighted work. Even in the prior cases, the copyrighted work was licensed, not sold, and the fact that the newer games use the company's servers does not change the fundamental nature of the argument in any meaningful way, because it does not change the fundamental nature of the act in question (the user modifying a game) or the technology in question (a device that helps the user do so) in any meaningful way.
What the company might be able to do is allow the technology to be sold, then sue the seller for damage to their reputation under the computer fraud and abuse act. This is, of course, entirely orthogonal to the copyright issue, which is settled law.
They should make it twice as hard to guess. Half the time, make it lisasuper.
Construction is only part of it. Yes, they often build commercial buildings cheaply, under the assumption that after twenty or thirty years, they will be so dated that nobody will want to rent them. But the bigger problem is that many building owners do minimal maintenance, under the assumption that any maintenance beyond what is critical — maintenance that could extend the buildings' life for decades — would probably not be cost-effective, because they'll be tearing them down to build something new after two to four decades anyway.
Which they don't own.
Oops. I have to correct myself. The IL campus is only 25 years old, not 40. So it still has some life left in it. It's the Mariani buildings that are almost 40 years old.
I'm basing this on the fact that buildings depreciate over 40 years for a reason. That's the expected life of the building. The buildings in question are now forty years old. So from a value perspective, these buildings are bulldozer fodder. They're not worth anything per square foot at that age. Anybody who buys them will be buying them solely for the right to lease the land for ~55 years. Mind you, if that lease is way below market value (and it may be), then the lease might have significant value, but either way, that value isn't in the building, but rather in the land use rights.
If this is their actual headquarters (Infinite Loop), then based on forty-year depreciation, a forty-year-old building is worth basically zero, so $200 seems about right. If this is Apple Park, then yeah, this is fraud.
Actually, it isn't entirely insane, assuming this is talking about the Infinite Loop campus. Apple does not own the land under the building. It is owned by Sobrato, the development company on the corner. Apple merely has a 100-year lease on it. So if Apple decided to sell the buildings, absent some agreement by the landowner to allow the lease to be transferred, they would not be able to do so. Arguably, then, the buildings have zero value beyond what the landowner is willing to give them for them.
And even if the landowner agreed to a lease transfer, the buildings would still only have value if somebody else wants them as-is. The problem is, IL1's lobby area had serious mold problems fifteen years ago, particularly on the upper floors. I can't imagine it has gotten any better since then. If Apple ever left, there's a nonzero chance that the next company would decide to tear those buildings down rather than fix them.
It could well be that the expected amount of money that they could get for transferring the lease would not significantly exceed the amount of money they would have to spend bulldozing the old campus to make it ready for whatever company would take it over.
Mind you, I do think that $200 is a gross underestimate, but if the city valued it at a billion dollars, that's a laughable overestimate. There's no way you'd get anywhere close to that for a bunch of forty-year-old buildings, no matter how much history they might have.
And given that the original 100-year lease is almost halfway up, and at the end of that 100 years, the buildings potentially become a giant teardown liability unless the lessee is willing to move them somewhere else, the value of those buildings is at least arguably going to go *negative* at some point.
So really, the only reasonable way to value the property is to determine how much Apple would have to pay to move the employees that are currently in the Loop to other, rented office space, multiplied over the expected remaining life of the building — maybe ten years on the high side. If we assume that they stopped doubling and tripling up in IL offices after Apple Park opened, that's probably only a couple of thousand people. And assume that any new space would be high-density, open plan office space at 175 square feet per employee. Assuming about $4 per square foot per month times ten years, that's about $168 million. At $8 per square foot for demolition times 850,000 square feet, the buildings themselves are a $6.8 million liability, so its value is really closer to $160 million. Seems like a much more plausible number than a billion, which would basically require assuming that Apple will continue using those buildings as-is for the remainder of the hundred-year lease.
Not really. Lidar basically just gives you a depth map. You can get the same thing from a pair of cameras, which Teslas have, at least on the front (the only direction that really matters that much). That doesn't require sci-fi-level strong AI; that's literally what the iPhone does every time you take a photo in portrait mode. It's basically a glorified version of an MPEG encoder's P-frame motion vector computation. It doesn't even require ML, much less strong AI. It is just a bunch of motion vector computation between two reference frames.
Just tell them that there's black gold (oil) in them thar hills, and they'll be ready to bomb it in under an hour. :-D
Slashdot posted the bug this morning. If I'm reading it right, the bug was reopened at 10:59 last night. So unless there's time travel involved or something.... :-D
I think most of us were pretty certain a long time ago that it would require a faster computer than what they were using. That said, if they're really able to get their machine learning model to process 2,000 frames per second with the new hardware, up from 200, they went from limping along at 25 fps per camera to ostensibly processing 250 fps per camera, which finally exceeds the actual frame rate of the camera hardware (60 fps).
I wouldn't call that "not working". I would call that grossly underestimating the compute power required for self-driving. Chances are, they'd have encountered the same problem if they had used LIDAR like Waymo. It isn't a panacea; it's just another data source.
It was introduced in 1974, but its popularity really took off in the 1980s. Interestingly, cancer incidence in the United States has been on a steady climb in the U.S. since about 1980. That increased incidence could entirely be a side effect of better detection, or Monsanto could be responsible for one of the biggest public health crises in our nation's history, or anything in between.
Trust me, not making sense never stopped the U.S. Government — particularly in the current administration.
And probably at least one who has a different opinion on whatever specific issue that person is complaining about. After all, although there are some moral absolutes, there are a lot more situations where different sets of morals conflict, such as the conflict between getting self-driving tech onto the roads sooner to save lives when drivers are half asleep versus delaying it until it is better than those drivers when they are awake.
If you're holding either one, you're doing it wrong. Step 1: Place the bag in such a way that the bottom lies flat, with the sides piled up around it. Step 2: Wait for the cat to think that it is a box and squeeze between the vertical walls. Step 3: Raise the sides of the bag in a quick vertical motion.
If this does not work, try putting a small amount of cat food on the bottom of the bag.
No, they're still abandoned. Just as a mother screaming "Don't you touch my child" from a thousand miles away when she hears a bear over a cell phone does not qualify as protection, neither does sending out C&Ds for copying a game that is unavailable commercially qualify as proof of non-abandonment.
I fully agree with the folks who say that there should be a maximum rest period for copyrighted works. If companies want to put something temporarily on hold to allow for demand to build up enough to re-release it, that's one thing, but putting distribution on hold indefinitely is quite another, and runs almost directly contrary to the copyright act's stated purpose of encouraging the creation of new works. If a work has been off the market for ten years, it should automatically be ceded to the public domain.
So no, don't move on. Write your congresspeople. Tell them that you think the copyright act needs to be amended to make redistribution of long-term-unavailable works per se legal. Tell them that statutory limits on how long a rights owner can prevent a work from being available on the market can result in opportunities for new businesses to form and make money off of rights that otherwise presumptively will never again be exercised in any meaningful way. Tell them that these limits contribute to the ability of others to create derivative works that can actually create long-term benefits for the original rightsholder. And tell them that allowing a company to simply bury a work and prevent any new people from ever being able to see it through abusing copyright is just plain dumb.
If enough people do that, we might just make something happen.
This. I predict a week later, these will get auto-deleted by a bot for being insufficiently noteworthy. And thus will begin the first AI-powered edit war.
That's what the nighttime detergent spray is for. It washes the oil off their feathers, and then they freeze to death. Then, you just have to deal with the yard full of dead birds.
ESPN almost certainly doesn't subsidize Disney. They pay almost $2 billion for pro football alone, which is likely more than the total budget for all Disney Channel shows put together.
That said, other NBC-Universal properties might be subsidizing Sci-Fi somewhat. And certainly, all those crappy shopping networks subsidize Sci-Fi by paying cable providers to carry them. (I refuse to call it SyFy.)
Mainly so that all the parents with little kids can pay fees for Disney's mostly low-cost programming, to subsidize ESPN so that people who watch sports don't have to pay $30 a game.
Dallas isn't land-locked. It can expand almost infinitely. The Bay Area has mountains on basically all sides, and the land is all in active use. Builders *are* building homes as fast as they can, but it is hard, because before they can build, they have to buy something smaller and tear it down. If the land under that condo costs 12 million dollars and can only realistically handle 30 units, the builder would lose almost half a million dollars selling it for $55K. That's what's happening in the Bay Area. If you want to live in Sacramento, you can probably get a condo for $55K. You'll just have to commute for two hours each way every day down a highway that goes through one of only a handful of mountain passes into the Bay Area.
First, you're wrong. Commercial property is covered under Prop 13, too. This is one of the serious problems with that law that many Californians would like to fix, because it would bring in a lot of much-needed revenue from big businesses.
Second, most of the new construction is mixed-use these days, with shopping on the first floor and housing on the upper floors. You're right that in some areas (San Francisco in particular), efforts to build more housing have sometimes been thwarted by government idiocy, but the bulk of the problem is actually caused by allowing a few businesses to become too big to fail.
Basically, there are three big employers: Apple, Google, and Facebook. These three employers have a very disproportionate impact on hiring.
The problem is that they are all close together. With no traffic, it takes about 14 minutes to get from Apple to Google, and ten minutes from there to Facebook. More importantly, traffic going to Google and Facebook are basically using highway 101, taking nearly consecutive exits. The peninsula is a nightmare as a result, traffic-wise. And more importantly, they are all three towards the northwest corner of the South Bay, relatively speaking. So any housing built towards the southeast (the only remaining direction for expansion, down a narrow valley) requires everyone to travel the same highway for most of their trip.
Because they are so close together, the housing within driving distance of those companies is untenable. And there is almost no business construction in areas like Gilroy where housing is more affordable, which otherwise could create reverse commutes and make the whole traffic situation more sane, because business leaders think that somehow they won't be able to get people who live in the South Bay to commute out of the South Bay. Fixing this would help somewhat.
Finally, as previously noted, the Bay Area is largely land-locked, with mountains in nearly every direction. So unlike other cities that just expand over time, the Bay Area really can't.